Weighted Predictive Video Coding for Luminance-Change Scenes

The patent applies parameter changes by introducing weighted prediction parameters (weight and offset) that dynamically adjust the prediction process based on luminance change detection. When luminance changes are detected in reference pictures, the system modifies the prediction parameters to compensate for these changes, thereby maintaining high coding efficiency in luminance change scenes while still utilizing inter-frame prediction to reduce network bandwidth strain.

The patent implements dynamics by making the weighted prediction parameters adaptive rather than fixed. The system detects luminance changes in reference pictures and dynamically adjusts the weight and offset parameters accordingly. This allows the prediction effect to be maintained across varying distances between current and reference frames, while still keeping the code stream transmission burden low by only transmitting parameters when necessary.

The patent applies local quality by differentiating the prediction mode usage based on local luminance change characteristics. Instead of uniformly applying explicit weighted prediction to all blocks, the system detects luminance changes locally in reference pictures and applies weighted prediction parameters selectively to regions where luminance changes are detected. This maintains high prediction effect where needed while minimizing unnecessary parameter transmission.

The patent applies segmentation by dividing the picture into slices and applying adaptive weighted prediction independently to each slice. This allows different weighted prediction parameters to be used for different slices, providing flexibility to handle diversified luminance change situations in different regions of the picture. The segmentation approach maintains data transmission efficiency by allowing independent processing of each slice without requiring global parameter transmission.